Both small and large plants are likely to produce staminate (male) flowers in a hermaphrodite lily

Abstract

Resource allocation to female and male function may vary among individual plants in species with variable sex expression. Size-dependent sex allocation has been proposed in hermaphrodites, in which female-biased allocation may increase with plant size. In many hermaphrodites with large floral displays, however, later-produced flowers tend to be functionally male. This paradoxical relationship between female and male function and plant size remains poorly understood. The subalpine lily Lilium lankongense has individuals of three sexual types: males with only staminate flowers, hermaphrodites with only perfect flowers, and andromonoecious plants with both perfect and staminate flowers. Here we tested theoretical predictions of size-dependent sex allocation in L. lankongense by measuring plant height and flower number of individuals of each sex at five field sites in the mountainous region of Shangri-La, southwestern China. To investigate variation in phenotypic gender, we identified sex expression of 457 individuals one year later. Our investigation showed that male plants, which usually produced one flower, were significantly smaller than andromonoecious and hermaphrodite plants. In addition, the total flower numbers of andromonoecious and hermaphrodite plants increased significantly with plant size. Large individuals were more likely to produce terminal staminate flowers, as there were more flowers in andromonoecious than in hermaphrodite individuals. Non-flowered plants were significantly smaller than flowering ones. Perfect flowers had significantly larger petals and pistils than staminate flowers, but they did not differ in dry weight of stamens. Our findings indicate that when plants are small, the less costly sex is favored, consistent with the ‘size-advantage hypothesis’. When plants are large, both female and male investments change isometrically, as later-produced flowers tend to be functionally male.